Scribbler
Unleash JavaScript's Power: Experiment with Ease Using Notebooks!
FreeOpen SourceCodeFramework13,271 impressions#3 of its week
Scribbler is an in-browser, no-backend notebook for running JavaScript AI, data science, and visualization.
- for
- Developers and data scientists who want client-side AI without servers.
- pricing
- free
- license
- MIT
Key features
- Zero-setup — Runs entirely in the browser; no install, login, or backend required.
- WebGPU acceleration — Uses WebGPU for near-native performance on LLMs, image generation and ML inference.
- Dynamic library loading — Import TensorFlow.js, ONNX Runtime, Transformers.js, Plotly, D3 and more from CDNs on demand.
- Interactive notebooks — Mix JavaScript, HTML, CSS and Markdown in live cells with inline output.
- Privacy-first — All computation stays on the device; no data leaves the browser.
- Share & collaborate — Export notebooks as .jsnb files, share via URL, or push/pull directly to GitHub.
- Self-hostable — Pure static files can be hosted on any web server, S3 bucket or GitHub Pages.
- Built-in AI models — Run Stable Diffusion, LLM chat, and TTS via WebNN and ONNX Runtime in the browser.
Use cases
- Generate images with Stable Diffusion directly in the browser.
- Chat with open-source LLMs locally for private text generation.
- Train and run ML models using TensorFlow.js or ONNX Runtime without a server.
- Create interactive data visualizations with Plotly or D3 inside notebooks.
- Share reproducible AI experiments via URL or GitHub for team collaboration.
Scribbler vs alternatives
| Best for | Client-side AI notebooks | LLM observability | AI coding assistant | Online data analysis | Server-side image processing |
|---|---|---|---|---|---|
| Pricing | Free | Subscription | Free | Free | Free |
| DevHunt upvotes | 13 | 89 | 63 | 41 | 95 |
| Launched | Jan 2023 | Jan 2023 | Mar 2024 | Aug 2024 | Feb 2024 |
- Scribbler vs Langfuse: Langfuse focuses on observability for LLM apps, not on in-browser notebooks.
- Scribbler vs Pieces Copilot: Pieces Copilot provides AI code suggestions, whereas Scribbler runs full AI models in the browser.
- Scribbler vs DataLine: DataLine is a SaaS for data analysis and visualization, not a client-side notebook platform.
- Scribbler vs imgproxy: imgproxy is a self-hosted image processing service, not an interactive AI notebook.
Scribbler FAQ
Do I need to install anything to use Scribbler?+
No. Scribbler runs completely in the browser; just open the URL and start coding.
Is my data sent to any server?+
All computation happens on the client device, so your data never leaves the browser.
Can I use external JavaScript libraries?+
Yes, you can dynamically import libraries like TensorFlow.js, ONNX Runtime, Plotly, D3, and others from CDNs.
How do I share a notebook with others?+
Save the notebook as a .jsnb file, share the URL, or push the notebook to a GitHub repository.
What hardware is required?+
A modern browser with WebGPU support; no dedicated GPU or backend infrastructure is needed.
Is Scribbler open source?+
Yes, the project is open source and can be self-hosted from the GitHub repository.
Summarized by DevHunt from scribbler.live · Sep 27, 2026. Details may change; check the official site.